3D Printer with Ballscrews

mendelevium

@mrehorstdmd
The idea was to build a formidable machine. my issue is the grinding noise/shift when it was tackling circles and thin walls. However I managed to lower the speed and the nose went down. I am concerned by shifts really and that grinding noise which I suspect the steppers having to make quick moves.

mrehorstdmd

@mendelevium That can be a problem. A ball screw is massive and if the motor doesn't have sufficient torque it can slip. Resonance can cause slipping problems too. I used a big NEMA-23 motor to drive the ball screw in my printer. I ultimately pulled the ball screw out and went back to belt drive.

CCS86

I don't see where you define the stepper mode, so maybe you are defaulting to Stealthchop and it's the switch to spread cycle that is messing you up.

Try forcing spread cycle mode: ie: M569 P0 D2

gallaghersart

I have been doing a bunch of research to add ball screws to my CoreXY X and Y. So I could still be wrong.

Steppers motors die off in power after @500RPM. Some math at end..

If screw is fixed on both ends you may get some warp from heat expansion when using aluminum frames. Less is steel is used.

Regardless the size of Screw relates to how fast it can spin before deforming.

12xx are good for Z.

2020/2025, 2520,2525 screws for X and Y. Requiring NEMA 24/34 minimum. Motors need to stop and move a lot of mass.

I was going to use 1616 for X and Y, but after doing math the screws themselves would only be good for @100mms before deforming.

So then we get to the motors that you would need to move 20xx/25xx ball screws.

Steppers just do not do it. Need Servos, and they are the costly.

All ya, 1616 ball screws should have 15mm rails min. 20xx screws should have 20mm rails. Dual blocks per rail.

So, I am now at a point that smaller rails with liner motors may be best.

Also, you need to have you ball screws perfect with your rails. Even 0.01mm will mess things up.

Hope that helps,
`mike

Some math,

Stepper max rpm 500 = 8.3333333 Rotations per second
5 =41.666666665 mms
10 =83.33333333
15 =124.999999995
16 =133.3333328
20 =166.66666666
25 =208.333333325
32 =266.666666656
40 =333.333332 mms

SERVO max rpm 3000
=50 rps /25 rps /10 rps
5 =250 mms
10 =500mms /250mms /100mms
15 =750
16 =800
20 =1000mms at 50rps /500mms at 25rps / *Goal 200mms at 10rps
25 =1250
32 =1600
40 =2000

mendelevium

@gallaghersart
Thanks for your input... I believe the screws are not deflecting in any way because the load in a 2d printing is not significant. The motors I am using can deal with it. I think the issue is that the motor has to reverse a lot and at speed since with a ballscrew, the motor spins more.
I have good experience with ballscrews and servos as I have around 16 cnc turning and machining centers. Somebody mentioned forcing spread cycle. I will gave a look and give it a try. I have other similar 3d printers I built and will resist going for the belt with this machine...

o_lampe

@mendelevium said in 3D Printer with Ballscrews:

that grinding noise which I suspect the steppers having to make quick moves.

I had to rebuild my 1204 ballscrews lately, and was surprised how dirty it was inside the ballnuts. The screws only had a few hours lifting Z-axis on their clock. I guess they appreciate some service after run-in period. ( at least my cheap robodigg ballscrews liked it)

mendelevium

@o_lampe
my ball screws have dedicated lubrication. I lubricate them using slide oil by Q8- NS68 slideway lubricant. Ballscrews need to have a seal to take the dirt off but you need to lubricate them thoroughly since the oil drives the dirt away. Avoid dust and fluff as well. My movements run over Hywin reciprocating ball slide-way and they are original ones. there is a lot of Chinese imitation rubbish. I will try to find some time and post some photos.

gallaghersart

Nema 17 motors can max out even with belts.

Moving x any y axis fast on 12xx ball screws with any size of NEMA 17 will not work. Just because they are not hot does not mean they are missing steps.

The best stepper motors die off at 500rpm. No name or cheap will be worse.

If screws are 1205 then your max stepper motor speed would be 41.666666665mms.

The smallest motors I would even try would be NEMA 23 4.+ amp.
nema 17 2.4amp for z axis balls crews should be fine no fast movement, but i still use nema 23 for Z.

mendelevium

@mendelevium said in 3D Printer with Ballscrews:

so these are my

these are the parameters to introduce spread cycle
M569 P0 D2
how can I modify the following config parameters?

; NOTE Drives S0 forward S1 reverse
M569 P0 S1 ; X physical drive 0 goes backwards
M569 P1 S0 ; Y physical drive 1 goes forwards
M569 P2 S1 ; Z physical drive 2 goes backwards
M569 P3 S0 ; E0 physical drive 3 goes backwards
M569 P4 S1 ; E1 physical drive 4 goes backwards

droftarts

@mendelevium from your config.g, it looks like you’re running a Duet 2? Duet 2 WiFi/ Ethernet does not have stealthchop, on spreadcycle. Send M115 and post response to confirm this and firmware version.

I would suspect it’s your jerk/instantaneous speed causing problems. M566 is set to X900 Y900, which is 15mm/s. With high steps per mm, possibly under-specced NEMA 17s turning heavy ball screws, and the mass of the carriages to move, I’d say this was a lot to ask. You have to treat the X and Y axes like a Z axis, which normally wouldn’t have a high jerk. Try M566 X300 Y300. It may also help to try the alternative jerk policy by setting P1 at the end of M566.

Ian

3DPMicro

I have done extensive testing with ball screws before venturing into my current build. You dont need large diameter. Diameter is a function of critical speed and load rating although that's not a factor with the light loads in a 3d printer. Your rails do not have to match your ball screw diameter. That's just the trend and mostly applys to mills.

I have ran these setups for many hours of simulated printing
https://youtu.be/KWK_djAidQs
https://youtu.be/g2fO7k3WzwU
1210 screws, solid mounts and solid couplers . Higher rpm gets the stepper out of mid band resonance and we can easily pump them full of high step rate. 20mm lead would be better. Theres an open builds that uses 60mm lead lead screws and prints beautifully. Several professional grade machines use ball screws. Most with servos instead of steppers though

mendelevium

@droftarts
FIRMWARE_NAME: RepRapFirmware for Duet 2 WiFi/Ethernet FIRMWARE_VERSION: 3.3beta1 ELECTRONICS: Duet WiFi 1.02 or later FIRMWARE_DATE: 2021-02-14 16:00:49

I am running the Larger Nema 17 motors
Nema 17 Bipolar 1.8deg 65Ncm (92oz.in) 2.1A 3.36V 42x42x60mm 4 Wires which can handle 2.4Amp @ 24V.
Before I had the smaller nema 17 which were not strong enough.

I do not believe/feel that the motors are under rated though you might say better than me in this. It involves a lot to mount for larger motors as i do not have much space to mount them but will consider everything, even changing the control board if it sorts the problems. Having said this, the crunching noise/shift comes when dealing with small 3mmholes and thin sections. Somebody suggested motors with a 20mm lead. these will certainly run at lower speeds. What about the M566 and P1. I am willing to give it a try.
Thanks

mendelevium

@3dpmicro
just looking around... where did you get these? I am beginning to realize that our CNC machining centers have the X and Y axis with large pitches... only the Z axis is with a small pitch which would make sense...

mendelevium

@mendelevium
I have reduced the jerk value considerably an have have obtained good results. IT is obvious that when the ratio of motor movement to axis movement is high as in the case of a ball-screw,the jerk is magnified. I have reduced the value of M566 from 900 to 120 and all the noise has vanished and there seems very little reduction in speed I have another tweak in the pipeline but will open a new post as it is not related exclusively to jerk issues